© Senckenberg Gesellschaft für Naturforschung, 2014.

213

72 (2): 213 – 219

25.7.2014

ISSN 1863-7221 (print) | eISSN 1864-8312 (online)

DNA barcoding of flat bugs (Hemiptera: Aradidae) with phylogenetic implications

Vasily V. Grebennikov 1 & Ernst Heiss 2

1 Canadian Food Inspection Agency, Ottawa Plant Laboratory, KW Neatby Bldg, 960 Carling Ave, Ottawa, Ontario K1A 0C6, Canada; Vasily V. Grebennikov [vasily.grebennikov@inspction.gc.ca] — 2 Tiroler Landesmuseum, Josef-Schraffl-Straße 2a, A-6020 Innsbruck, Austria; Ernst Heiss [aradus@aon.at]

Accepted 17.vi.2014. Published online at www.senckenberg.de/arthropod-systematics on 18.vii.2014.

AbstractThis work introduces an open online library of Aradidae DNA barcodes, specimen images and geographical data with intent of promot-ing further community-based DNA barcoding of flat bugs. We report the results of an attempt to DNA barcode 191 dry specimens of the flat bugs representing all 8 extant subfamilies of the Aradidae. 145 sequences > 300 nt in length were obtained (76% success rate; all can be seen in a publicly accessible dataset “World Aradidae”, doi: dx.doi.org/10.5883/DS-WAHG), 98 of them representing 55 species and 29 genera were > 500 nt (51%). These 98 sequences were combined with 10 Heteroptera-Trichophora outgroup taxa into an aligned matrix of 658 positions and variously analysed to obtain the first DNA-based phylogeny for the Aradidae, which was our second goal. Aradidae and five of its subfamilies (Aradinae, Aneurinae, Mezirinae, Isoderminae, Calisiinae) were recovered as monophyletic, although the lat-ter two were markedly underrepresented. The subfamily Carventinae was consistently polyphyletic. Isoderminae + Prosympiestinae was recovered as a clade, while Aradinae formed the sister group to the rest of the family in three of four analyses. The genus Mezira appeared polyphyletic while the genus Neuroctenus was paraphyletic with respect to Ctenoneurus.

Key wordsAneurinae, Aradinae, Calisiinae, Carventinae, Chinamyersiinae, Isoderminae, Mezirinae, Prosympiestinae, Termitaphididae, Pentatomo-morpha, CO1.

1. Introduction

The nearly cosmopolitan terrestrial heteropteran family Aradidae, colloquially known as “flat bugs”, contains at least 1,798 markedly depressed medium-sized dark-ish species arranged in at least 211 genera, as per the somewhat outdated catalog by Kormilev & Froeschner (1987). Since then at least 56 new genera and 168 new species have been described, of them 35 genera and 88 species by the second author. Flat bugs are thought to be mycophagous, but very little of their biology is known (leschen & Taylor 1987). The majority of Aradidae in temperate zone ecosystems are winged and are normally found under bark, while most tropical species are wing-

less and inhabit forest leaf litter. For years the latter spec-imens were thought to be nymphs, until miller (1938) described the first exclusively apterous genus Chelono­coris (Fig. 1) from Malaysia. Since the pivotal volume of Usinger & maTsUda (1959), eight extant subfamilies have been consistently recognized. Among them, Meziri-nae is by far the largest, embracing more than half of all Aradidae species; the other being Aneurinae, Aradinae, Calisiinae, Carventinae, Chinamyersiinae, Isoderminae and Prosympiestinae. Only New Zealand and Australia have representatives of all subfamilies (larivière & la rochelle 2006). The ninth subfamily, Archaearadi-

Grebennikov & Heiss: Aradidae barcoding and phylogeny

214

nae, is extinct and contains two monotypic genera from Burmese Amber of the mid- and upper-Cretaceous, re-spectively (heiss & Poinar 2012). The mid-Cretaceous Cretopiesma suukyiae Grimaldi & Engel, 2008 from Myanmar possibly represents another subfamily (cas-sis & schUh 2010), while at least 52 other extinct aradid species have been attributed to four of the eight extant subfamilies (references to more than half of them can be found in KaUlFUss et al. 2011). The phylogenetic placement of Aradidae is less con-troversial than that for many other bug families. The fam-ily is either the sister taxon to Termitaphididae (schUh & slaTer 1995; henry 1997) or more likely paraphyletic (grimaldi & engel 2005, 2008; cassis & schUh 2010) with respect to that family. The latter includes 13 exclu-sively termitophilous, blind and wingless small-bodied species (Fig. 1) arranged in two genera and known from all main tropical regions, as well as from Mexican and Dominican Miocene amber (Poinar & heiss 2011). Ow-ing to the rarity of specimens, Termitaphididae eluded phylogenetic analyses until the morphological analysis in cassis & schUh (2010) consistently placed it deeply nested within the strongly supported Aradidae: the single representative used in their analysis, Termitaradus guia­nae Morrison, 1923, was placed as sister to the Aradi-nae-Chinamyersiinae. Among other characters, the two families share a flattened body and elongate mouthparts coiled inside the head; the latter character is a seemingly complex and unique synapomorphy. The superfamily Aradoidea (or Aradidae sensu lato, i.e. including Termit-aphididae) is consistently placed as sister to the rest of Pentatomomorpha, which is often termed “Trichophora”, reflecting the presence of trichobothria on the ventral sur-face of the abdomen (see henry 1997; cassis & schUh 2010). Pentatomomorpha is one of the seven heteropteran infraorders (others being Enicocephalomorpha, Dipsoco-romorpha, Gerromorpha, Nepomorpha, Leptopodomor-pha and Cimicomorpha; see WeiraUch & schUh 2011). Three publications have addressed relationships with-in the Aradidae using morphological characters. vásár-helyi (1987) and grozeva & Kerzhner (1992) provided intuitive topologies with the eight extant subfamilies as terminals. cassis & schUh (2010) performed a cladistic analysis based on 34 terminals (including eight extant Aradidae) and 78 characters and included two critically important terminals: a species of Termitaphididae and the fossil Cretopiesma suukyiae. The latter work was, however, not designed to resolve relationships within the Aradidae, but to test the placement of the fossil and that of Termitaphididae. No DNA-based topology has ever been proposed for the Aradidae. We used the BOLD online platform to develop the first open library of Aradidae DNA barcodes (heberT et al. 2003) in order to facilitate future research on this group similarly as it was done with many other animal groups (see, for example, hogg et al. 2009 for New Zea-land Trichoptera, deWaard et al. 2011 for British Co-lum bia Geometridae, sPelda et al. 2011 for Bavarian Myriapoda). While most projects have examined speci-

mens from a limited geographic area, we removed this constraint by analyzing specimens from the highly re-presentative collection of Aradidae assembled by Ernst Heiss. Until now the family was poorly represented in the Barcode of Life Database (BOLD) with only 15 CO1 sequences other than ours. Our second goal was to utilize the sequence data, to the extent possible, to propose the first DNA-based Aradidae phylogenetic hypothesis and to compare it with those advanced previously on the ba-sis of morphological characters.

2. Material and methods

Specimen origin, depository and choice of taxa. All specimens barcoded in this project came from, and are deposited in, the research collection of Ernst Heiss in Innsbruck, Austria. They were collected by various peo-ple under different circumstances and were stored as dry pinned specimens for 8 – 12 years before being used for DNA extraction. Specimens and species were chosen by EH to proportionally represent each of the eight extant Aradidae subfamilies and some of the most diversified genera without any geographical bias, although Ecuador (18 records), Malaysia (15), India (13), Madagascar (13) and Austria (13) were the five most strongly represented countries. A single leg of a dry specimen was removed for DNA extraction, as customary for medium-sized in-sects and other arthropods.

Aradidae genetic barcodes. In June 2014 the project “Barcoding World Aradidae” contained 191 records, each represented by an imaged specimen, 59 with geo-referenced data. These specimens belonged to 73 iden-tified species, plus a few species not identified beyond the genus, representing all eight currently recognized subfamilies of the Aradidae, although four (Calisiinae, Prosympiestinae, Isoderminae, Chinamyersiinae) were represented by only 2, 2, 1 and 1 sequences, respectively. Chinamyersiinae was particularly poorly represented as only one 325 nt sequence was obtained from the Aus-tralian Kumaressa scutellata Monteith, 1966, so it was excluded from subsequent analyses. To maximize the amplification success rate, two am-plification attempts were made: the first one with the PCR primers LepF2_t1/LepR1 targeting the entire stand-ard 658 nt “barcoding” region of the CO1 mtDNA gene (heberT et al. 2003), and the second one with a cocktail of two PCR primer sets LepF2_t1/MHemR and MHemF/LepR1 targeting two shorter and partly overlapping se-quences of the same 658 nt region. Among the total of 191 specimens, 46 failed to amplify and contained no as-sociated DNA data, while the remaining 145 were each represented by sequences > 300 nt (GenBank accession numbers KF809495 – KF809639). Information on 145

215

ARTHROPOD SYSTEMATICS & PHYLOGENY — 72 (2) 2014

sequenced Aradidae voucher specimens with their digital images and all relevant data such as primers and original chromatograms, was deposited in the Barcode Of Life Database in the publicly accessible dataset “World Ara-didae”, doi: dx.doi.org/10.5883/DS-WAHG. Among the 145 sequences, 98 were > 500 nt and represented 55 spe-cies and 29 genera of seven subfamilies (see Table 1).

Matrix construction and outgroup. The analysed data-set contained 108 sequences with the minimal/maximal lengths of 518/658 nt having no indels and was unam-biguously aligned. It included all 98 barcode-compliant Aradidae records, plus 10 outgroup sequences. The latter were chosen among the publicly available “barcoding” sequences of “Trichophora”, the putative sister-group of Aradoidea, representing eight families and all four super-families: (BOLD process ID followed by GenBank ac-cession number, whenever available, are given in brack-ets): Pentatomoidea: Pentatomidae: Euschistus servus (HCNCS389-09 HQ105678) and Podisus serieventris (HCNC785-09 HQ106267); Scutelleridae: Vanduzeeina balli (HCNC778-09 HQ106454) and Homaemus aenei­frons (HCNC197-09 HQ105749). Pyrrhocoroidea: Lar gi- dae: Physopelta australis (MAMTF1450-12); Pyr rho co-ridae: Probergrothius sp. (GBMHH432-10 GU247509). Lygaeoidea: Lygaeidae Lygaeus kalmii (HCNC755-09 HQ105855); Piesmatidae: Piesma cinereum (RBI-NA1761-13). Coreoidea: Coreidae Leptoglossus occi­den talis (HCNCS251-09 HQ105828); Rhopalidae: Rho­pa lus tigrinus (HCNCS316-09 HQ106310).

DNA substitution model and topology building. The search for a substitution model required for two out of four analyses (BI and ML; see below) was performed us-ing MEGA5 (TamUra et al. 2011). Four different topo-logy-building analyses were implemented and are as fol-lows (in order of the increasing computation complexi-ty): the Neighbor-Joining method (NJ) using uncorrected p-distances with 1000 replicates bootstrapping was per-formed using MEGA5; the Maximum Parsimony method (MP) with 1000 replicates bootstrapping was performed using MEGA5; the Maximum Likelihood method (ML) using the GTR+G+I model (see Results) with 1000 re-plicates bootstrapping was performed using MEGA5; the Bayesian Inference method (BI) using the commands “lset nst = 6 rates = gamma” and three million generations with the default burn-in using MrBayes 3.2.2. (ronqUisT et al. 2012). The NJ analysis, being a not phylogenetic approach, is provided here only for comparative pur-poses since it is the most common and fast-performing algorithm widely used with the DNA barcoding data.

Tree manipulation and visualization. The resulting BI and three majority consensus topologies obtained by bootstrapping the dataset by using NJ, MP and ML meth-ods were visualized in FigTree (rambaUT 2013) for sub-sequent comparison. All four obtained topologies were first examined unrooted. Since in all four of them all Aradidae and all non-Aradidae taxa were always form-

ing two sister clusters, the root was forced between them and the bootstrap support value for this internode was taken as that indicating monophyly of Aradidae. The ma-jority consensus topology from the ML analysis showing only branches with > 10% support was then exported as a Windows Enhanced Metafile (.emf), the outgroup man-ually deleted, and the tree enhanced in CorelDraw and then Photoshop (Fig. 1). Adult habitus images illustrat-ing terminal taxa are specimens actually genetically bar-coded and can be seen at higher resolution in the online BOLD project.

Phylogenetic limitations. Our phylogenetic interpreta-tions presented below should be taken with caution con-sidering their unavoidable limitations. The size of the matrix could have been much greater, either in number of nucleotides or in number of terminals. Two critically important taxa were not represented: the obligatory in-quilinous family Termitaphididae and the south-Pacific subfamily Chinamyersiinae (see monTeiTh 1980; the single available sequence of Kumaressa scutellata was judged too short to be analysed). Three oligotypic sub-families Prosympiestinae, Isoderminae and Calisiinae were represented by a single specimen, species and ge-nus, respectively, thus depriving us of an opportunity to test their monophyly. Perhaps even more acute was the limitation imposed by use of only the “barcoding” region of the CO1 mitochondrial gene. The latter is generally considered too quickly saturating for adequate represen-tation evolutionary events as old as the early radiation of Aradidae, which likely dates back to at least the mid-Cretaceous (120 – 90 Myr; see heiss & Poinar 2012), and, therefore, the reported results should be treated with much caution.

3. Results and discussion

Statistics. The generalised time-reversible substitution model (GTR) with gamma distributed rate heterogeny (G = 0.7120) and inferred proportion of invariable sites (I = 37.9976%) was of the best overall fit. The Bayesian Inference method after 3,000,000 generations had the standard deviation of split frequencies 0.009363. The BI tree and majority rule bootstrapping consensus trees rep-resenting different analytical method (NJ, MP and ML) consistently recovered monophyletic Aradidae with rela-tively similar internal branching patterns summarized on Table 1.

DNA barcoding of Aradidae. The method performed adequately with our Aradidae samples, even though only half of the submitted samples resulted in full se-quence-recovery after trying three sets of generalized insect primers. Only 76% (145 of 191 samples submit-

Grebennikov & Heiss: Aradidae barcoding and phylogeny

216

ted) yielded DNA sequences > 300 nt, and only 51% (98 samples) were > 500 nt (= barcode compliant). The success ratio of 51% is much lower than the 95% typi-cal of most arthropod groups. The moderate success in recovery likely reflects the fact that the specimens were relatively old (8 – 12 years) and not specifically collected or stored with DNA work in mind. On the other hand, however, our recovery rate is indeed a result of interest, exactly as it was based on the traditional way of insect collecting and pinning, not targeted on DNA work, using specimens 8 – 12 years dead. Seen from this side, the rate is satisfyingly high. Even though not designed to test the minimal intra- and inter- specific/generic distances, our results correspond with those which have tested capacity of the method to discriminate species in other arthropod groups, including Heteroptera (JUng et al. 2011; ParK et al. 2011). We conclude that DNA barcoding is likely to perform well for species discrimination in the Aradidae, particularly when sequences are supplemented by other data such as specimen images and geo-references, all supported by the on-line BOLD platform.

Phylogeny of Aradidae. The monophyly of Aradidae has never been questioned based on any available evi-

dence, and, therefore, the consistent recovery of the fam-ily as a clade with each of the four DNA-based analyses is not surprising (clade A on Fig. 1). It is interesting that such a well-supported group (cassis & schUh 2010) was only strongly recovered by BI (99%), but only weakly in NJ, MP and ML (51 – 53%; see Table 1). This fact reflects the well-known limitation of using a short fragment of the fast-evolving CO1 gene to recover early evolutionary events (maddison 2012). Besides recovering the Aradidae as a clade, the larg-est subfamily Mezirinae, represented in the analysis by 72% of all ingroup genera, is the only diversified group with more than three genera that was consistently recov-ered in the analysis. The monophyly of the Mezirinae (clade J on Fig. 1) is arguably our most important result, since it corroborates continuous recognition of the sub-family embracing over 50% of all flat bug genera and species. Statistical support for Mezirinae was markedly higher than that for Aradidae (see Table 1) and with less discrepancy between BI and other three analytical ap-proaches. Four other of the seven analysed subfamilies we re-covered as clades: Aneurinae, Aradinae, Calisiinae and Isoderminae. Among these, Calisiinae (clade I on Fig. 1)

Table 1. Select groups of Aradidae (with the number of analysed specimens, species and genera), if recovered as clades (clade lettering as in Fig. 1) and their statistical support on four consensus trees, each representing one of the following methods: Neighbor-Joining (NJ; non-phylogenetic and used for comparative purposes only), Maximum Parsimony (MP), Maximum Likelihood (ML) and Bayesian Inference (BI). Support values are percentages of 1000 bootstraps (NJ, MP, ML; clades supported with less than 10% are ignored and collapsed) or posterior probabilities (BI; clades supported with less than 50% posterior probabilities are ignored and collapsed), multiplied by 100. A zero value indicates the taxonomic group was not recovered as a clade. Taxonomic abbreviations: Bir: Biroana, Neu: Neuroctenus, Cte: Ctenoneurus, Sty: Stysaptera, Atr: Atractocoris, Mal: Malgasyaptera, Per: Pericaptera.

  Specimens Species Genera Clade NJ MP ML BI

Aradidae 98 54 29 A 53 52 51 99

Aradinae 14 10 1 B 40 33 46 53

Aradidae minus Aradinae 84 44 28 C 17 0 17 84

PRO + ISO + Bir 4 3 3 D 43 40 45 100

PRO + ISO 3 2 2 E 78 79 79 0

Isoderminae (ISO) 2 1 1 F 100 100 100 100

Aneurinae 8 4 2 G 33 29 35 91

CAL + Sty + Atr + Mal + Per 10 7 5 H 15 38 15 0

Calisiinae 2 2 1 I 39 58 34 100

Mezirinae 62 30 19 J 71 81 72 100

Neu + Cte + MezSp 24 8 3 K 45 36 45 85

Carventinae 13 8 6 none 0 0 0 0

Prosympiestinae (PRO) 1 1 1 none n/a n/a n/a n/a

→ Fig. 1. Phylogenetic relations of flat bugs (Hemiptera: Aradidae) inferred from the “barcode” segment of the CO1 gene using the Maxi-mum Likelihood bootstrap majority consensus from 1000 replicates. Support values are indicated above the respective branches; branches with less than 10% support are collapsed. Lettering on branches (A – K) highlights clades discussed in the text. The vertical dimension of the terminals is proportional to the number of specimens analysed. Habitus images are denoted by abbreviated genus and species letters on the same level with the terminal and are not to scale. No representatives of the Termitaphididae and Chinamyersiinae (illustrated to the top left of the tree) were included in the analysis. Branch length is not to scale.

217

ARTHROPOD SYSTEMATICS & PHYLOGENY — 72 (2) 2014

Grebennikov & Heiss: Aradidae barcoding and phylogeny

218

were represented by only two records of two congeneric species, Aradinae (clade B on Fig. 1) by only 10 con-generic species, Aneurinae (clade G on Fig. 1) by only four species from two genera, and Isoderminae (clade F on Fig. 1) by only two records of a single species. Each of these four subfamilies was, therefore, significantly underrepresented in the analysis and their monophyly could not be rigorously tested. Prosympiestinae was represented in our analysis by a single record, thus no evidence at all could be obtained on its monophyly. Car-ventinae was split in two (BI) or three (NJ, MP, ML, Fig. 1) separate clades, each consistently placed on the tree. Thus the genus Biroana formed a clade with Prosympi-estinae and Isoderminae (clade D on Fig. 1), while four other Carventinae genera (Stysaptera, Atractocoris, Mal­gasyaptera and Pericartaptera) grouped with Calisiinae (clade H on Fig. 1). Two taxonomic groups of lower rank were consist-ently recovered as non-monophyletic. The genus Mezira, the type genus of the most speciose subfamily, was con-sistently recovered in two subunits (Fig. 1). The genus Neuroctenus (clade K on Fig. 1) was found paraphyletic with respect to Ctenoneurus (Fig. 1). The inconsisten-cies discovered for the three latter genera are not surpris-ing, since they all urgently need a revision, particularly with respect to numerous species in the Oriental Region, which, as stated by Kormilev (1971), cannot be assuredly assigned to either Neuroctenus or Mezira. Three recent studies bear on subfamily arrangement of Aradidae. vásárhelyi (1987) suggested monophyletic Chinamyersiinae to be the sister to the rest consisting of: (Aradinae + Calisiinae) + (Isoderminae + (Prosym-piestinae + (Aneurinae + (Mezirinae + Carventinae)))). grozeva & Kerzhner (1992) accepted (Aradinae + Cal-isiinae) as sister to Chinamyersiinae-Tretocorini, and the entire group as sister to Chinamyersiinae-Chinamy-ersiini + ((Prosympiestinae + Isoderminae) + (Aneuri-nae + (Mezirinae + Carventinae))). The subfamily-level arrangement of Aradidae was not the primary goal of cassis & schUh (2010) and their topologies were not dogmatic in this respect, however three points of their results relevant to our work are: (1) Prosympiestinae and Isoderminae do form a clade, (2) this clade is sister to the rest of the family (with or without a representative of non-monophyletic Chinamyersiinae) and (3) Aradinae is never sister to the rest of the family. Most of our results cannot be adequately compared with those above, except that in three among four analyses we recovered Aradi-nae as sister to the rest of the family (clade C on Fig. 1, except for MP) and Prosympiestinae and Isoderminae forming a clade (clade E on Fig. 1, except for BI). At present we attribute all such discrepancies to the lack of knowledge and relative immaturity of phylogenetic Ara-didae research and, therefore, have to be content with the results pending further more sizable and focussed ef-forts. In conclusion, it is appropriate to speculate on what next might be the most logical step in revealing the phylogenetic history of Aradidae. A morphology-based

analysis similar in scope and implementation to those of other comparably diversified family-level clades, like the Omaliine Group of rove beetles (neTWon & Thayer 1995), or Reduviidae assassin bugs (WeiraUch 2008), is long overdue. Analysis of larger DNA matrices is another obvious approach. Supplementing future topologies with geographical and, particularly, biological information will markedly add to their informative value and eventu-ally lead to a revised family classification at all levels. Reliable species identification and revived interest to flat bugs will be necessary to do so, and the herein released online DNA barcode and image collection is expected to stimulate future phylogenetic work. Until then the pre-sent status quo eight subfamily arrangement of visionary Usinger & maTsUda (1959) seems an adequate approxi-mation.

4. Acknowledgments

Karen McLachlan Hamilton (Ottawa, Canada) critically read an early draft of this paper. Sequence analysis was carried out at the Canadian Centre for DNA Barcoding (CCDB) at the University of Guelph with support provided by the Government of Canada through Genome Canada and the Ontario Genomics Institute for the International Barcode of Life project. We thank Paul Hebert (Guelph, Canada) for overseeing the sequence analyses and for comments on our manuscript.

5. References

cassis G., schUh R.T. 2010. Systematic methods, fossils, and re-la tionships within Heteroptera (Insecta). – Cladistics 26: 262 – 280.

deWaard J.R., heberT P.D.N., hUmble L.M. 2011. A comprehen-sive DNA barcode library for the looper moths (Lepidoptera: Geometridae) of British Columbia, Canada. – PLoS ONE 6: e18290. doi:10.1371/journal.pone.0018290.

grimaldi D., engel M.S. 2005. Evolution of the Insects. – Cam-bridge University Press, New York, xv + 755 pp.

grimaldi D.A., engel M.S. 2008. A termite bug in Early Miocene amber of the Dominican Republic (Hemiptera, Termitaphidi-dae). – American Museum Novitates 3619: 1 – 10.

grozeva S.M., Kerzhner I.M. 1992. On the phylogeny of aradid sub families (Heteroptera, Aradidae). – Acta Zoologica Hunga-rica 38: 199 – 205.

heberT P.D.N., cyWinsKa A., ball S.L., deWaard J.R. 2003. Bio-logical identifications through DNA barcodes. – Proceedings of the Royal Society of London B 270: 313 – 321.

heiss E., Poinar G.O. 2012. New Aradidae in Mesozoic Burmese amber (Hemiptera, Heteroptera). – Annalen des Naturhistori-schen Museums Wien, Serie A 114: 307 – 316.

henry T.J. 1997. Phylogenetic analysis of family groups within the infraorder Pentatomomorpha (Hemiptera: Heteroptera), with

219

ARTHROPOD SYSTEMATICS & PHYLOGENY — 72 (2) 2014

em phasis on the Lygaeoidea. – Annals of the Entomological Society of America 90: 275 – 301.

hogg I.D., smiTh B.J., banKs J.C., heberT P.D. 2009. Testing use of mitochondrial COI sequences for the identification and phy-logenetic analysis of New Zealand caddisflies (Trichoptera). – New Zealand Journal of Marine and Freshwater Research 43: 1137 – 1146.

JUng S., dUWal R.K., lee S. 2011. CO1 barcoding of true bugs (In-secta, Heteromera). – Molecular Ecology Resources 11: 266 – 270.

KaUlFUss U., WaPPler T., heiss E., larivière M.-C. 2011. Aneurus sp. from the early Miocene Foulden Maar, New Zealand: the first Southern Hemisphere record of fossil Aradidae (Insecta: Hemiptera: Heteroptera). – Journal of the Royal Society of New Zealand 41: 279 – 285.

Kormilev N.A. 1971. Mezirinae of the Oriental Region and South Pacific (Hemiptera-Heteroptera: Aradidae). – Pacific Insects Monograph 26: 1 – 165.

Kormilev N.A., Froeschner R.C. 1987. Flat bugs of the world: a synonymic list (Heteroptera: Aradidae). – Entomography 5: 1 – 246.

larivière M.-C., larochelle A. 2006. An overview of flat bug genera (Hemiptera, Aradidae) from New Zealand, with con-siderations of faunal diversification and affinities. In: rabiTsch W. (ed.), Hug the bug: for love of true bugs: Festschrift zum 70. Ge burstag von Ernst Heiss. – Plöchl-Druck, Freistadt; Denisia 9: 181 – 214.

leschen R.A.B., Taylor S.J. 1987. Notes on the biology of Aradus robustus (Hemiptera: Aradidae). – Entomological News 98: 183 – 185.

maddison D.R. 2012. Phylogeny of Bembidion and related ground beetles (Coleoptera: Carabidae: Trechinae: Bembidiini: Bembi-diina). – Molecular Phylogenetics and Evolution 63: 533 – 576.

miller N.C.E. 1938. A new subfamily of Malaysian Dysodiidae (Rhynchota). – Annals and Magazine of Natural History, Se-ries 11 1: 498 – 510.

monTeiTh G.B. 1980. Relationships of the genera of Chinamyersi-i nae, with description of a relict species from mountains of North Queensland (Hemiptera: Heteroptera: Aradidae). – Pa-cific Insects 21: 275 – 285.

neWTon A.F., Thayer M.K. 1995. Protopselaphinae new subfamily for Protopselaphus new genus from Malaysia, with a phyloge-netic analysis and review of the Omaliine Group of Staphyli-nidae including Pselaphidae (Coleoptera). Pp. 219 – 320 in: PaKalUK J., Ślipiński S.A. (eds), Biology, Phylogeny, and Clas-sification of Coleoptera: Papers Celebrating the 80th Birthday of Roy A. Crowson. – Muzeum i Instytut Zoologii PAN, War-szawa.

ParK D.-S., FooTTiT R., maW E., heberT P.D.N. 2011. Barcoding bugs: DNA-based identification of the true bugs (Insecta: He-miptera: Heteroptera). – PLoS ONE 6: e18749. doi:10.1371/journal.pone.0018749.

Poinar G., heiss E. 2011. New Termitaphididae and Aradidae (He-miptera) in Mexican and Dominican amber. – Palaeodiversity 4: 51 – 62.

rambaUT A. 2013. FigTree, Version 1.4. Available at: http://tree.bio.ed.ac.uk/software/figtree/.

ronqUisT F., TeslenKo M., van der marK P., ayres D.L., darling A., höhna S., largeT B., liU L., sUchard M.A., hUelsenbecK

J.P. 2012. MrBayes 3.2: Efficient Bayesian phylogenetic infer-ence and model choice across a large model space. – System-atic Biology 61: 539 – 542.

sPelda J., reiP H.S., oliveira-biener U., melzer R.L. 2011. Bar-coding Fauna Bavarica: Myriapoda – a contribution to DNA se quence-based identifications of centipedes and millipedes (Chi lopoda, Diplopoda). – ZooKeys 156: 123 – 139.

TamUra K., PeTerson D., PeTerson N., sTecher G., nei M., KU mar S. 2011. MEGA5: molecular evolutionary genetics ana lysis us-ing Maximum Likelihood, Evolutionary Distance, and Maxi-mum Parsimony Methods. – Molecular Biology and Evo lu tion 28: 2731 – 2739.

Usinger R.L., maTsUda R. 1959. Classification of the Aradidae. – British Museum, London, 410 pp.

vásárhelyi T. 1987. On the relationships of the eight aradid sub-families (Heteroptera). – Acta Zoologica Hungarica 33: 263 – 267.

WeiraUch C. 2008. Cladistic analysis of Reduviidae (Heteroptera: Cimicomorpha) based on morphological characters. – System-atic Entomology 33: 229 – 274.

WeiraUch C., schUh R.T. 2011. Systematics and evolution of He-teroptera: 25 years of progress. – Annual Review of Entomo-logy 56: 487 – 510.